The future’s bright, the future’s artificial

Nations such as Wales are putting innovation at the forefront to help rebalance systems that keep people well…

Artificial intelligence (AI) and machine learning have the potential to transform the future of healthcare as we start the road to recovery from the Covid-19 pandemic’s initial impact. With its integration into both medical workflows and hospital systems, we can expect to see dramatic changes in both patient health outcomes and hospital operational efficiency.

Healthcare systems globally face numerous, large-scale system challenges to meet their citizens’ evolving needs. For the last 20 years, healthcare costs have outstripped fiscal growths across more economically developed countries. Current systems are unsustainable, with clinical and adjacent staff facing predictable demand pressures. We simply cannot continue investing in the same service models of the past. Recognising the challenges and need for transformative change, governments are embracing long-term plans to deliver health, wellness, and prevention-focused integrated systems of care. Nations such as Wales are putting innovation at the forefront to help rebalance systems that keep people well, living in the community and out of hospital for longer.

Digital innovation can play a key part in achieving this. We remain in the early stages of understanding, developing, and embedding AI into our health and care systems, but it can and will help deliver transformational change if stakeholders work together.

Transforming imaging and diagnostics

AI diagnostic systems can be programmed to automate data detection and interpretation. This can support resource management, a critical need as the UK radiologist workforce was 33 per cent short staffed in 2020¹. Delivering accurate diagnosis efficiently and non-invasively can significantly support the push towards preventative healthcare through early-stage intervention.

Take medical imaging, where AI can analyse scans to support triaging. If integrated seamlessly into imaging workflows, it can increase efficiencies. This frees up time to allow clinical teams to focus on more complex clinical projects and diagnostics and helps patients receive treatment more quickly.

An example of market-ready technology to support this is AI lung nodule identification for early-stage cancer detection. This uses deep learning, which enables artificial neural networks to adapt and learn from vast amounts of data. It is designed to be fully integrated with existing worklflows to detect, quantify, assess growth, and classify nodules in both routine clinical practice and screening programmes - streamlining processes and reducing patient waiting times.

Fully integrating such technologies across multiple diagnostic fields needs large-scale transformation. For example, detecting cancer often uses a range of diagnosis and treatment types across many medical centres with many clinical parameters. Moreover, current AI diagnostic tools cannot account for patient history or multiple diseases or conditions, both of which come with increasingly vast volumes of data.

Overcoming this challenge centres on aggregating data and developing multi-modality AI systems. Here, we need AI platforms that can collect and analyse diagnostic data covering all stages of the cancer pipeline and with inter-operability between all involved systems. One company or an individual cannot do this. We need strong partnerships between academics, pathology and imaging experts, healthcare providers, patient advocates and developers.

Data-driven precision medicine

Diagnostics can also be transformed through machine learning-powered precision medicine. Clinicians have access to a huge array of treatment options for conditions, but certain treatments have no guarantee of working across all patients – potentially causing harm and wasting valuable resources. Precision medicine is underpinned by vast swathes of biological and patient datasets that capture variation in genes, functions, and the environment. Machine learning methods can help to automate such approaches when in the clinic, rapidly sorting through and identifying patterns in genetic data and information on treatments, interventions, and medicines. While in early phases of research and development, it could help pave the way to a preventative future – where accurate disease predication and personalised treatments are deployed by clinical staff with ease.

Harnessing the power of AI

To advance AI adoption, innovators need financial support and infrastructure to deliver digital healthcare innovation in healthcare settings. Projects such as the NHSX AI in Health and Care Award competition can provide essential stimulus and networks. Ecosystems must also continue to

grow and be supported to accelerate AI innovation. Multidisciplinary collaboration is essential for such development and delivery. Solutions must reflect the diverse expertise from the experience of clinicians, developers, life sciences industry, academia, and the patient.

Complex ethical and social issues also need to be considered and regulatory frameworks built to safeguard public confidence. This will help to ensure this technology is used effectively to maximise patient benefits.

For more info watch the ‘Digital and AI in Wales’ video on lshubwales.com/our-priorities/digital-and-artificial-intelligence-ai

References

1 Royal College of Radiologists (2020). Clinical radiology UK workforce census 2020 report.

Author: 

Cari-Anne Quinn is CEO of Life Sciences Hub Wales lshubwales.com